Running a GPAQ survey by post

The postal version of GPAQ has been designed to give more general practice information. Although it does ask about the patient’s “usual” doctor, it should not be used for individual assessments of doctors – for this you should use the “consultation” version of GPAQ.

The following are some useful points to consider if you are administering your own GPAQ survey by post.

Take time to make sure the sample is correct

This will pay dividends in the end and enhance the response rate. Make sure the sample is legitimate and up-to-date and that the sampling strategy you use is clear. For example, make sure patients you send the questionnaire to are in the correct age range (16 years or older) and that they are still registered with the practice. It also pays to check if any patients on the sample list have died.

Get the sample size right

Quality & Outcomes Framework guidance recommends that 25 questionnaires per 1000 registered patients on the practice's list should be returned. However, in terms of individual assessment we would still recommend that 50 questionnaires are collected per doctor for statistical reliability, even though many practices with smaller list sizes (e.g. less than 2000 patients) could collect less than 50 questionnaires per doctor.

Write a short covering letter

Make sure it is clear that the practice has given full support of the survey and that doctors and nurses will not be able to identify individual patients from the responses. Also make it clear that whatever views the patient expresses, these will not affect their future care. Give the phone number of someone in the practice to phone if they have questions.

Use postal reminders

After around two weeks, send out reminder notices to patients who have not yet returned the questionnaire. You should aim for a response rate of 60%, and you will almost certainly need at least one reminder to achieve this. In order to know which patients have responded to a postal survey, you will also have to put a patient identification number on the top of each questionnaire, linked to a list of patients you have sent the questionnaire to. Because you will need to use a patient identification number, it is very important to make sure it is clear that the doctors and nurses will not be able to identify individual patients from their responses. Patient identification numbers can also be used to enable checking of data entry at a later point (we recommend checking 10% of the data that is entered).

Do not underestimate how long it will take

It takes a significant amount of time to administer the survey (including sending reminders) and to enter the data (unless this is being done by professional data entry clerks, in which case do not underestimate the cost). Administering the survey is probably the hardest and most time consuming part.

Go to downloads to download the questionnaire to send by post.